Adaptive Spot-Guided Transformer for Consistent Local Feature Matching

Deep Space Exploration Lab, 3 China Academy of Space Technology (a)Reference (b)Linear Attention (e)Matching Result (d)Ours (c)Vanilla Attention Figure 1. The visualization of the cross attention heatmaps and matching results. We sample two similar adjacent points in the reference image (a), marked with green and red. (b) are two heatmaps of the linear cross attention in LoFTR [50] when green and red pixels are queries. (c) are two heatmaps obtained from the vanilla cross attention. (d) are two heatmaps generated by our spot-guided attention. (e) are the comparison of the final matching results produced by LoFTR [50] (top) and our method (down).